This document presents a project on liver segmentation in biomedical applications using deep learning techniques. The objectives are to study liver segmentation algorithms, collect and analyze datasets, and design a web application for liver segmentation using a hybrid ResUNet model combining ResNet and UNet. The methodology involves pre-processing CT images, adapting algorithms, and examining results using quality metrics. Hardware and software requirements for implementing the proposed project using Python, VS Code, ResUNet framework are readily available.
1. DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING
LIVER SEGMENTATION IN BIOMEDICALAPPLICATION
USING DEEP LEARNING TECHNIQUE
SDM INSTITUTE OF TECHNOLOGY
Presented By
Namana Hegde
Samreen
Sneha S J
Abdul Imad
Under the guidance of
Supritha P O
Assistant Professor
Department of CSE
SDMIT, Ujire
TEAM ID:
2022CSEPT29
2. INTRODUCTION
• Liver segmentation process is aimed to divide the pixels of the image
depending on the certain criteria.
• An accurate and robust tumours segmentation method is needed for
effective diagnosis and treatment. Consequently, premature analysis and
precise review of tumours are critical. But, due to the minor noticeable
deviations among healthy tissues and tumoral ones the detection of liver
and its tumours is challenging.
• To detect the tumours a good and repeatable technique would be an
advantage. In order to evade various and consecutive segmentations a
clinically motivated approach must work for various tumour types which is
a quite difficult task at the same time.
• Because of complexity of liver shapes and variable liver sizes among
patients the segmentation of liver from medical images is very difficult and
also due to low contrast between liver and surrounding organs like
stomach, pancreas, kidneys, muscles.
2
3. LITERATURE SURVEY
1.Automatic 3-D Skeleton-Based Segmentation of Liver
Vessels from MRI and CT for Couinaud Representation
Published in: 2018 25th IEEE International Conference on Image Processing (ICIP)
This study proposes an automatic 3-D skeleton-based segmentation for liver vessels. We identify
vascular network to propose a segmentation of Couinaud anatomical segments. Experimental
results show that the proposed method can segment liver vessels on both modalities with faithful
3-D representations.
2. Liver Segmentation in CT Images Using Deep Neural
Networks
Published in: 2020 28th Iranian Conference on Electrical Engineering (ICEE)
Automatically extracting the liver from CT or MR images due to its heterogeneous shape and
proximity to other organs is a challenging task. This research presents an algorithm to perform a
detailed liver segmentation. In this algorithm, images are first classified with a classification
network to be separated into the liver included and non-liver included classes, then the class
containing the liver are analyzed with the segmentation network.
3
4. LITERATURE SURVEY
3. Liver segmentation: A survey of the state-of-the-art
2017 Sudan Conference on Computer Science and Information Technology
(SCCSIT)
The preoperative such as planning to understand the complex internal structure of the liver and accurately
localize the liver surface and its tumors; there are numerous algorithms proposed to do the automatic liver
segmentation. In this survey paper, we analyze critically some of different published works for liver
segmentation algorithms since 2007 till 2016. This paper also compares and contrasts the methods,
datasets, results and limitation for each work. A comprehensive comparative analysis is conducted.
4. Automatic Liver Segmentation Using Multi-plane Integrated Fully
Convolutional Neural Networks
Published in: 2018 IEEE International Conference on Bioinformatics and
Biomedicine (BIBM)
Automatic liver segmentation in 3D medical images is essential in many clinical applications, such as
pathological diagnosis of surgical planning, postoperative assessment and hepatic diseases. Our network
uses multiple layers of dilated convolution filters to replace traditional ones. Our method outperformed
other state-of-the-art methods with an average Dice score of 96.7% on the segmentation results of liver,
which only used a single framework without any pre-processing operation on it.
4
5. LITERATURE SURVEY
5.Liver Tumor Segmentation and Classification: A Systematic
Review
M. Rela, N. R. Suryakari and P. R. Reddy, "Liver Tumor Segmentation
and Classification: A Systematic Review," 2020 IEEE-HYDCON
The traditional methods used for detection of the tumor are time-consuming,
give the error in detection, and these methods require experts to analyze the
tumor. Hence automatic and integrated methods are required to use instead of
traditional methods. Liver tumor segmentation from CT image is very
important to analyze the liver function, pathological and anatomical study of
the liver. It is also important for the diagnosis of disease. This paper discusses
various methods for early detection of liver tumors and also discussed the
merits and demerits of these methods.
5
6. PROBLEM STATEMENT
According to the most recent estimates from global cancer statistics for 2020,
liver cancer is the ninth most common cancer in women. Segmenting the liver
is difficult, and segmenting the tumor from the liver adds some difficulty. After
a sample of liver tissue is taken, imaging tests, such as magnetic resonance
imaging (MRI), computer tomography (CT), and ultrasound (US), are used to
segment the liver and liver tumor.
Due to overlapping intensity and variability in the position and shape of soft
tissues, segmentation of the liver and tumor from computed abdominal
tomography images based on shade gray or shapes is undesirable. This study
proposed a more efficient method for segmenting liver and tumors from CT
image volumes using a hybrid ResUNet model, combining the ResNet and
UNet models to address this gap.
6
7. OBJECTIVES
The objectives of the proposed project are as follows:
1. To study and analyze the related work through study of liver segmentation.
2. To study and identify suitable Algorithms for liver segmentation.
3. To collect and analyze the data set for projects.
4. To design and develop the web application for liver segmentation.
7
8. METHODOLOGY
8
Step 1: The first is to perform pre-processing operators to enhance the CT images.
Step 2: The second stage is the adaptation of algorithms to the problem, considering
equal conditions.
Step 3: The third and final stage is to examine the results by using quality metrics and
predict sample images.
•
9. Requirements and feasibility
9
Sl. No Software Specification
1. VS Code Code Editor
2. Python Python 3 and above
3. ResUNet Framework
• Hardware Requirements
• Software Requirements
• All the hardware and software components required for implementing the proposed project is
readily available. Hence project can be carried out within given time constraint and budget.
12. REFERENCES
• World health organization.
http://www.who.int/mediacentre/factsheets/fs297/en
• World Health Organization. World Cancer Report. 2021
https://www.who.int/news-room/fact-sheets/ detail/cancer
• Key Statistics about Liver Cancer. 2022
https://www.cancer.org/cancer/livercancer/about/whatiskeystatistics.html#:~:text=T
he%20American%20Cancer%20Society’s%20estimates,will%20die%20of%20thes
e%20cancers
12